Journal of Student Research 2021

Time-Series Analysis of Wave Elections

77

Time-Series Analysis of Wave Elections

James Evans 1 Senior, Applied Math and Computer Science Faculty Advisor: Dr. Tyler Skorczewski

Abstract A wave election is commonly defined as an election in which one party makes major gains. However, there is little consensus on what constitutes such gains. A wave election could also be thought of as an election in which the electorate voted in a fundamentally different way than preceding elections. In this study, we create a quantitative definition using ideas found in time-series analysis. We apply segmentation algorithms to election data to find the moments when these fundamental shifts in the electoral mood occur and build an objective definition of a wave election by comparing these results with past accepted wave election occurrences. Introduction What is a wave election? This question is what our study seeks to answer. The term “Wave Election” has risen to popularity [1] in recent years. The term also increases in use around the time of elections. Figure 1 shows the Google Trends report for the term “Wave Election.” Not only does its usage spike around the time of the election, but these spikes have also been increasing [1] in magnitude in recent years. This increased usage drives a need for a consistent definition to avoid misunderstanding and misuse of the term. According to Oldham and Smith [2], people use wave elections as a measure of political change. This is because they often correspond with dissatisfaction [3] among the general populous for a party. Thus, a wave can be considered a sign of a fundamental change in the mood of the voting populous.

Figure 1: Google Trends[1] report for the search term ``Wave election`` 2004 to present.

1

James is a McNair Scholar and a member of the Honors College.